Available online at www.sciencedirect.

com

ScienceDirect

Functional ultrasound neuroimaging: a review of the

preclinical and clinical state of the art
Thomas Deffieux1,2, Charlie Demene1,2, Mathieu Pernot1,2 and
Mickael Tanter1,2

In the last decade, ultrasound imaging has gained new portability, machine costs, maintenance and accessibility
capabilities and produced new insights in the field of [3]. Furthermore, limited sensitivity and patient stress in
neuroscience. The development of new concepts, such as the confined and noisy magnet are of concern [4].
ultrafast ultrasound, has enhanced Doppler sensitivity by
orders of magnitude and has paved the way for ultrasonic Position emission tomography (PET) is another func-
functional neuroimaging. In this review, we position ultrasound tional imaging technique that uses injected radioactive
in the field of neuroimaging and discuss how it complements and biologically active tracers, such as fluorodeoxyglucose
current tools available to neurobiologists and clinicians. (FDG), to image brain molecular processes; this enables a
visualization of the consumption of glucose related to
Addresses brain metabolism [5]. Positron Emission Tomography
1
Institut Langevin, CNRS, ESPCI Paris, Inserm, PSL Research computes a 3D reconstruction of the positron-emitting
University, Paris, France
2
radionuclides concentration from the pairs of gamma rays
Inserm Technology Research Accelerator in Biomedical Ultrasound, they emit indirectly. Positron Emission Tomography
Paris, France
(PET) is a powerful and highly sensitive nuclear imaging
Corresponding author: Tanter, Mickael ([email protected]) modality but requires dedicated and secured facilities due
to the radioactive nature of the tracers used. PET suffers
from a poor spatial resolution and must be combined with
Current Opinion in Neurobiology 2018, 50:128–135
a complementary imaging modality, such as MRI or
This review comes from a themed issue on Neurotechnologies Computed Tomography (CT), for anatomical imaging.
Edited by Anikeeva and Luo
Finally, optical imaging techniques are fast and provide
excellent spatiotemporal resolution. These can be used to
image blood flow variations at different scales: from the
https://doi.org/10.1016/j.conb.2018.02.001 submillimeter resolution with intrinsic optical imaging to
0959-4388/ã 2018 Published by Elsevier Ltd. a micrometer resolution with two-photon microscopy.
Calcium imaging, which requires a genetically engi-
neered animal model, allows spectacular views of neuron
firing with high spatiotemporal resolution for surface
imaging or in transparent samples such as zebrafish [6].
In that case, fluorescent molecules are used as calcium
Introduction indicators as they respond to the binding of Ca2+ ions by
The longstanding quest to visualize the brain’s complex changing their fluorescence properties. However, the
organization and interconnections has led to major dis- limited penetration depth usually confines it to the sur-
coveries in modern neuroscience and neurology. Beyond face of the cortex. Optics can be used at higher depths at
anatomical imaging, the advance of technologies in the the cost of a substantial loss of spatial resolution due to
last decades has enabled the first dynamic images of the light scattering in tissues [7]. For example, functional
brain in action [1]. near-infrared spectroscopy, which measures the level of
blood oxygenation through the absorption of diffused
The gold standard of functional brain imaging is undoubt- near infrared light at different wavelengths, can be
edly functional magnetic resonance imaging (fMRI) to applied on neonates or young children through the skull
measure blood oxygen level-dependent (BOLD) signals but is limited to cortex imaging with centimeter spatial
on high field MRI scanners [2]. Since deoxygenated resolution [8].
hemoglobin is paramagnetic while oxygenated hemoglo-
bin is diamagnetic, blood deoxygenation introduces a Fundamental barriers of ultrasound imaging in terms of
magnetic signal variation, the BOLD signal. The BOLD temporal and spatial resolutions have recently been bro-
signal is indirectly related to neuronal activation through ken [9,10]. These paradigm changes led to functional
neurovascular coupling. fMRI is available for human use brain information using neurovascular coupling (Figure 1).
and is non-invasive and non-ionizing. Its primary draw- Interestingly, such functional ultrasound (fUS) neuroim-
backs both for preclinical and clinical imaging are its aging is a stand-alone ultrasound technique that provides

Current Opinion in Neurobiology 2018, 50:128–135 www.sciencedirect.com

 Functional ultrasound neuroimaging: a review Deffieux et al. 129

Figure 1

Imaging
plane
Ultrasonic
Probe
Fontanel

Brain connectomics

Neonate
neuroimaging

3D Functional imaging

Peroperative
Superresolution
neuroimaging
Angiography

Portable scanner

EEG compatibility
Awake &
Freely moving
Animal models

Current Opinion in Neurobiology

The main applications and features of functional ultrasound (fUS) imaging. fUS imaging provides (i) a compatibility with a wide range of animal
models for preclinical studies, (ii) the ability to image awake and freely moving animals, (iii) possibility to combine with super-resolution ultrasound
localization microscopy, (iv) possible extension to 3D imaging, (v) functional connectivity mapping for brain connectomics, (vi) translation to clinical
neuroimaging in human neonates or (vii) peroperative neuroimaging during brain surgery and (viii) EEG compatibility for EEG-fUS recordings.

high sensitivity imaging of cerebral blood volume (CBV) is a pertinent parameter for functional imaging that is
changes for whole brain imaging without contrast already used by other modalities such as intrinsic optical
agents [11]. imaging or CBV-weighted fMRI. The spatiotemporal
extent of CBV response was extensively studied
Ultrasound-based functional imaging thanks to these techniques [12], and the spatial resolution
techniques of sensory-evoked CBV response can go down to one
Neurovascular coupling cortical column (100 mm). Temporally, the CBV
Similar to all neurofunctional imaging techniques based impulse response function was measured to typically
on metabolic or hemodynamics measurements, functional start at 0.3 s and peak at 1 s in response to ultrashort
ultrasound is limited by the spatiotemporal features of stimuli (300 ms), which is much slower than the underly-
neurovascular coupling as it measures CBV changes. CBV ing electrical activity.

www.sciencedirect.com Current Opinion in Neurobiology 2018, 50:128–135

130 Neurotechnologies

Functional transcranial Doppler (fTCD) fUS research platforms require custom sequences pro-
Ultrasound Doppler imaging has long been used to obtain gramming, dedicated high-performance GPU beamform-
basic functional measurements of brain activity using ing software with a high data transfer rate (several GBytes
blood flow. In functional transcranial Doppler sonogra- per second) and miniature high-frequency ultrasound
phy, a low-frequency (1–3 MHz) transducer is used probes to perform live fUS imaging. Future commercial
through the temporal bone window with a conventional implementations through specialized hardware and soft-
pulse Doppler mode to estimate blood flow at a single ware should enable fUS to rapidly expand in utility for the
focal location. The temporal profile of blood velocity is neuroscience community.
usually acquired in main large arteries such as the middle
cerebral artery (MCA). The peak velocity is measured
and compared between rest and task conditions or Functional photoacoustic computed tomography
between right and left sides when studying lateralization Using laser devices, the photoacoustic effect can be
[13]. However, due to its restriction to global effects in leveraged to enable molecular imaging of optical contrast
large vessels and single point measurements, fTCD lacks at the ultrasound resolution [20]. The laser is used to
true neuroimaging capabilities. illuminate the brain while the strong light absorption by
red blood cells creates a sharp localized temperature
Power Doppler and contrast ultrasound imaging increase, which in turn generates ultrasonic waveforms.
Power Doppler is a Doppler sequence that measures the This latter signal is recorded by a tomographic ultrasound
ultrasonic energy backscattered from red blood cells in array and was recently able to detect sensory-evoked
each pixel of the image. Power Doppler provides no activation in rats [21]. Usually requiring a complex setup
information on blood velocity but is proportional to blood with a high-power pulsed laser, water-filled tank and
volume within the pixel. However, conventional power ultrasound ring, recent progress has nonetheless been
Doppler imaging lacks sensitivity to detect small arter- achieved toward wearable probes for awake imaging
ioles/venules and thus is unable to provide local neuro- [21] and integrated whole body solutions [20]. Due to
functional information through neurovascular coupling its sensitivity to optical contrast agents, photoacoustic
[14]. Adding acoustic contrast agents (microbubbles) to imaging could one day use calcium probes to perform
the blood stream boosts the sensitivity of conventional calcium imaging at the ultrasound resolution and depth
power Doppler imaging and enables the detection of [22] even in small animal models (Figure 2).
coarse brain activation in various areas of the brain [15].
Preclinical applications of functional
Ultra-fast ultrasound and fUS imaging ultrasound neuroimaging
fUS imaging relies on ultrafast imaging scanners [9] able In 2011, fUS technique was demonstrated for imaging of
to acquire images at thousands of frames per second, activation of the barrel cortex following whisker stimula-
thus boosting the power Doppler signal-to-noise ratio tion in rats [11]. CBV images are then correlated with the
(typically over 50-fold) without any contrast agents stimulus pattern, and activated regions are overlaid with a
[14]. Instead of the line per line acquisition of conven- state-of-the-art brain atlas (Figure 3d). In the same paper,
tional ultrasound devices, ultra-fast ultrasound takes spatiotemporal dynamics of epileptiform seizures were
advantage of successive tilted plane wave transmissions filmed showing cortical spreading depression propagating
that are afterward coherently compounded to form images throughout the whole brain (Figure 3e).
at high frame rates. The sensitivity was recently even
further improved using multiple plane wave transmis-
sions [16] and advanced spatiotemporal clutter filters for Chronic imaging and noninvasiveness
better discrimination between low blood flow and tissue Several animal preparations were proposed for fUS. Ini-
motion [17]. tially, a craniotomy was performed before the experi-
ments and an experiment was performed under isoflurane
This signal boost enables the sensitivity required to map anesthesia [11]. For chronic imaging, a thinned skull
subtle blood variations in small arterioles (down to 1 mm/ procedure was later proposed, allowing several weeks
s) related to neuronal activity, whereas conventional of imaging [19,23]. In Sieu et al. [19], a thin layer of
power Doppler is limited to imaging major cerebral polymethylpentene (PMP) with a high ultrasonic trans-
arteries (several cm/s) [9]. fUS neuroimaging has a typical mission coefficient and very low ultrasonic absorption was
50–200 mm spatial resolution depending on the ultra- used to replace the bone. This allowed maintaining
sound frequency used [14]. It features a temporal resolu- excellent imaging quality along with implanted electro-
tion in the tens of milliseconds, can image the full depth des in the animals for more than a year. In Rungta
of the brain and can provide 3D angiography [18]. fUS et al. [24], a polished PMP layer transparent to both
imaging requires no calibration and nearly no setup time. visible light (>93%; haze <5%) and in the high UV-range
It uses miniaturized probes to enable whole-brain imag- (>300 nm) and ultrasound was used to combine fUS with
ing in awake and freely moving rodents [19]. optogenetics and two-photon imaging.

Current Opinion in Neurobiology 2018, 50:128–135 www.sciencedirect.com

 Functional ultrasound neuroimaging: a review Deffieux et al. 131

Figure 2

High
High
Optical Implanted
MEG
Imaging EEG

Functional
Temporal Resolution

Ultrasound

Surface FNIRS
EEG

FMRI
y
lit
bi
rta

PET
Po

Low Whole Brain Imaging Local Brain Imaging

Low
Low Spatial Resolution High
Current Opinion in Neurobiology

Main brain functional imaging techniques on a three-axis chart (temporal resolution, spatial resolution, portability). Techniques were separated
between local and whole-brain imaging. Functional ultrasound fills a gap between whole brain imaging and microscopy, as well as between fMRI
and Optics.

Tiran et al. demonstrated the feasibility of performing arrest and resuscitation was also investigated in a rabbit
functional imaging directly through the skull in mice and model [40].
young rats until 35th postnatal day (P35) [25]. Using
microbubble contrast agents, Errico et al. [26] further
demonstrated non-invasive fUS through the adult rat Functional imaging in awake animals
skull. Regarding anesthetics, a mix of ketamine/Domitor One key aspect of fUS is its portability. First, the setup
is often preferred to isoflurane as inhalable anesthetics can be easily moved into animal facilities, which is in
can cause an increase in cerebral blood flow caused by contrast to fMRI scanners. More importantly, the probe
reduced cerebrovascular resistance. weight allows for experiments in awake and freely mov-
ing subjects. This removes the compounding effects of
anesthesia and facilitates behavior studies.
Full brain accessibility, 3D imaging and optogenetics
Extensive studies have been performed for somato-sen- The first studies were performed on freely moving rats
sorial stimulations. Olfactory stimulation in rats was dem- using miniaturized probes and head implants. In Sieu
onstrated in Osmanski et al. [27], who exposed the ani- et al. [19], a hybrid setup with a motorized miniaturized
mals to two different molecules while imaging the probe, implanted electrodes and acoustically transparent
anterior piriform cortex, which is a challenging structure skull prosthesis was proposed. This setup allowed func-
to image using fMRI due to nearby air cavities. Visual tional activation studies in awake animals in combination
stimulation and retinotopies were also investigated in with electrophysiological recordings anywhere in the
Gesnik et al. [28] who applied varying visual patterns brain using implanted electrodes. The method was used
to study the activation of the superior colliculus, the to study spatiotemporal initiation and the evolution of
lateral geniculate nuclei, and the visual cortex. Such seizures in epileptic GAERS rat models in conjunction
visual stimulation was later reproduced in mice, starling with EEG recordings. In the same work, rats running in a
birds and very recently in 3D in pigeons using the same maze were imaged to study locomotion by correlating the
approach and isoflurane anesthesia [29]. In Osmanski intra-hippocampal EEG theta band with vascular flow
et al., fUS was used in conjunction with optical fibers patterns. Merging real time data from fUS and another
to demonstrate the vasodilating effect of light in geneti- complimentary modality, such as electrophysiological
cally intact mice [24]. Massive hyperemia after cardiac recording, opens numerous possibilities [30]. Urban

www.sciencedirect.com Current Opinion in Neurobiology 2018, 50:128–135

132 Neurotechnologies

Figure 3

Freely moving setups Ultrasensible power Doppler Functional neuroimaging Functional connectivity

(a) (c) (g)

(d)

S1Sh-L
S1HL-L
M1-L
M2-L
RSD-L
(e) (f) RSGc-L
RSGc-R
RSD-R
(b) M2-R
M1-R
S1HL-R
S1Sh-R
Hip-L
Hip-R
Thal-L
Thal-R

RSGc-R
RSGc-L

S1HL-R
S1Sh-R
S1HL-L
S1Sh-L

RSD-R
RSD-L

Thal-R
Thal-L
Hip-R
M2-R
M1-R

Hip-L
M1-L
M2-L
fUS anesthesized fUS awake Transcranial Thin-skull Craniotomy
Mouse [24] [25] [24,25] [24]
Rat [11,14,18,23,25–27,28,34] [19,23] Microbubbles [26] [15,23] [11,14,18,19,26,27]
Young rats <P35 [25,34]

Bird [29] [29]

Rabbit [40] (post cardiac arrest resuscitation) [40]

Ferret [32,39] [32,39]
Primate

Current Opinion in Neurobiology

Preclinical applications of fUS imaging. Setups for awake rats [19] (a) or mice [25] (b). (c) Ultra high-sensitivity Doppler allows whole-brain
imaging in rats [14]. (d) Hyperemia induced by whiskers stimulation in the barrel cortex and in the ventral posterior medial nucleus [11]. (e)
Propagation of an epileptiform seizure in the rat brain [11]. (f) 3D reconstruction of the activated visual system of an anesthetized rat [28]. (g)
Resting state connectivity matrix in an anesthetized rat. The anti-diagonal represents the interhemispheric functional coupling [33]. Lower panel:
different protocols and animal models tested.

et al. [31] used rats running within a corridor and demon- Brain connectomics
strated visual and whisker stimulations with real time Similarly to resting state fMRI, functional connectivity
activation inside the cortex or in subcortical structures was measured in rats by Osmanski et al. [33] using fUS,
such as the thalamus. yielding high resolution (100 mm) brain connectivity
matrices. This protocol was later used to study the func-
In Tiran et al. [19], even smaller transcranial probes tional connectivity in rat pups following a low protein diet
permitted experiments in freely moving mice. Barrel in their mother [34]. The diet induced intrauterine
cortex activation was demonstrated following manual growth restriction and a loss of corpus callosum myelina-
whisker stimulation. The auditory tract of awake ferrets tion detected on the connectivity matrices obtained by
was imaged when listening to different high-pitch fUS. It could thus become a very convenient tool to study
tones, enabling the mapping of the tonotopy of auditory preclinical brain connectomics for drug development and
cortex and thalamic nuclei with a 100 mm resolution screening.
[32,39].
Clinical application of fUS imaging
Finally, the supplementary high field (SEF) of trained Early in-human neurofunctional studies using ultrasound
primates performing saccades and antisaccades visual were based on the use of only two ultrasonic transducers
tasks was very recently demonstrated and showed a placed on each temple window (Figure 4a), enabling the
30% increase of blood flow in parts of the SEF during assessment of cerebral blood flow differences in the left
tasks compared to baseline (Dizeux A, et al., IEEE and right Median cerebral artery (MCA) during a later-
Ultrasonics Symposium, 1948–5727, Washington, DC, alized cognitive task. This was used to show that language
Nov. 2017). functions were less predominantly localized in the left

Current Opinion in Neurobiology 2018, 50:128–135 www.sciencedirect.com

 Functional ultrasound neuroimaging: a review Deffieux et al. 133

Figure 4

TCD Lateralization Studies Human Neonate fonctional imaging Human Intraoperative fonctional imaging
(a) (c) (d) (g)
Imaging
Inferior
plane
Anterior Posterior
Ultrasonic Superior
Probe
Fontanel

(h) (i)
(b) (e) 80 (f) Stimulus
0 1
60 ON
ROIs Seizure
1

Correlation coefficient
60 40
UfD signal [%]

UfD signal [%]

Number of Subjects

Depth (mm)
20
40 OFF
0 110 220
0 Time (s)
20
–20

Le 0 –40
ft
La t
ng Righ –60
ua
ge ess –20 30 0
dn 500 0 110 220 0 25
Ri n de
gh Ha 0min + Time (s) Width (mm)
t
L eft

Current Opinion in Neurobiology

Clinical neuroimaging using ultrasound. (a) Conventional transcranial Doppler imaging. (b) Degree of language lateralization in relation to
handedness (from Ref. [28]). (c) Ultrasonic probe for neurofunctional Ultrasound on a neonate fontanel. (d) Ultra high-sensitivity Doppler image
acquired through the fontanel. (e) Relative changes in CBV occurring after 500 s. (f) 2D mapping of the hyperemia during and after the seizure
[29]. (g) Ultrasonic probe positioned on the brain during open-skull surgery. (h) Temporal profiles during a motor task in the motor cortex (top) and
surrounding areas (bottom). (i) Correlation maps with the stimulus pattern (red line in h) depicts the implicated motor cortex [31].

hemisphere in families with a high rate of left handedness optical and fMRI techniques. Ultrasound can be used
(Figure 4b in [35]). fUS took this to the next level by non-invasively on neonates through the fontanel window
translating 2D functional imaging into clinical setting in and through the skull in mice. In other configurations
neuropediatry and neurosurgery. In both cases, the use of with thicker skulls, contrast agents can be injected to
a natural anatomic window, that is, the transfontanellar ensure noninvasiveness; alternatively, a special skull
window, (Figure 4c,d) or craniotomy (Figure 4g) provides preparation such as craniotomy or thinned skull proce-
high-quality brain images in humans. In preterm neo- dure is possible. The technique is highly transportable
nates, fUS non-invasively imaged brain activity during (being in a form factor close to that of a conventional
different sleep phases and seizures [36], showing local ultrasound machine) and can thus be moved to patient
variations in epileptic activity (Figure 4e,f). Its temporal rooms, into wet labs or animal facilities without any
and spatial resolution unveiled a slow propagative phe- recalibration. The technique was demonstrated to be
nomenon after epileptic activity, corroborating previous compatible with simultaneous electrophysiology record-
in vitro brain slices experiments [37]. In neurosurgery, the ings, optogenetics and even PET scans. Moreover, the
technique was able to achieve real-time cortical func- dimensions of probes and cables have been reduced to
tional mapping (Figure 4h) during tumor resection [38], enable experiments in awake and freely moving rodents.
permitting in-depth delineation of cognitive areas and The technique currently suffers from limitations: tran-
avoiding the removal of functionally essential structures scranial imaging without contrast agents remains an issue
(Figure 4i). In the future, transtemporal imaging could except for in mice or niche applications. Additionally,
potentially enable fully non-invasive fUS on human experiments in freely moving mice need further minia-
adults and could extend the clinical domain of fUS. turization, and the technique remains primarily two-
dimensional. Preclinical fUS imaging is envisioned to
Conclusion and perspectives spread widely across the neurobiology labs thanks to
As a neuroimaging modality, functional ultrasound offers optimized and easy-to-use setups. This wide clinical
a unique combination of spatiotemporal resolution, sen- dissemination will prompt researchers to address the issue
sitivity, portability, and features that domains of bridge of transcranial ultrasonic propagation in the adult brain.

www.sciencedirect.com Current Opinion in Neurobiology 2018, 50:128–135

134 Neurotechnologies

Conflict of interest statement imaging increases signal-to-noise ratio in ultrafast ultrasound

imaging. Phys Med Biol 2015, 60:8549-8566.
TD, MP, MT are co-founders and shareholders of
17. Demené C, Deffieux T, Pernot M, Osmanski B-F, Biran V,
Iconeus company.  Gennisson J-L, Sieu L-A, Bergel A, Franqui S, Correas JM et al.:
Spatiotemporal clutter filtering of ultrafast ultrasound data
highly increases doppler and fultrasound sensitivity. IEEE
Acknowledgements Trans Med Imaging 2015, 34:2271-2285.
This work was supported by a research grant from the European Research The authors introduce a signal processing method to discriminate tissue
Council under the European Union’s Seventh Framework Program (FP7/ and blood flow motion in ultrafast ultrasonic raw data. This clutter filter
strongly outperforms all conventional filters used in ultrasound imaging. It
2007-2013)/ERC Advanced grant agreement n 339244-FUSIMAGINE and
further enhances the sensitivity of ultrafast Doppler to detect very low
by the Inserm Technology Research Accelerator in Biomedical Ultrasound. blood flows even in the presence of motion artifacts.
18. Demené C, Tiran E, Sieu LA, Bergel A, Gennisson JL, Pernot M,
References and recommended reading  Deffieux T, Cohen I, Tanter M: 4D microvascular imaging based
Papers of particular interest, published within the period of review, on ultrafast Doppler tomography. Neuroimage 2016 http://dx.
have been highlighted as: doi.org/10.1016/j.neuroimage.2015.11.014.
Using a tomographic acquisition of ultrasonic raw data, the authors
 of special interest introduce a methodology to image the dynamics of the whole brain
 of outstanding interest vasculature in rodents with an isotropic 100  100  100 mm3 voxel
and a millisecond temporal resolution. The authors also demonstrate
the very high sensitivity of fUS imaging to very small blood flow speeds
1. Friston KJ: Modalities, modes, and models in functional (down to 1 mm/s) involved in very small brain vessels.
neuroimaging. Science 2009, 326:399-403.
2. Logothetis NK: What we can do and what we cannot do with 19. Sieu L-A, Bergel A, Tiran E, Deffieux T, Pernot M, Gennisson J-L,
 Tanter M, Cohen I: EEG and functional ultrasound imaging in
fMRI. Nature 2008, 453:869-878.
mobile rats. Nat Methods 2015, 12:831-834.
3. Van der Kolk AJ, Hendrikse J, Zwanenburg J, Visser F, Luijten PR: Using miniaturized and motorized ultrasonic probes and EEG recording
Clinical applications of 7 T MRI in the brain. Eur J Radiol 2013, electrodes, this publication demonstrates for the first time the ability of
82:708-718. functional ultrasound coupled with electrophysiology to record the whole
brain activity in awake and freely moving animals. They record epileptic
4. Tomasi D, Caparelli EC, Chang L, Ernst T: fMRI-acoustic noise seizures in awake GAERS rats (a genetic model of absence seizures) and
alters brain activation during working memory tasks. freely animals moving in mazes.
Neuroimage 2005, 27:377-386.
20. Li L, Zhu L, Ma C, Lin L, Yao J, Wang L, Maslov K, Zhang R,
5. Phelps ME: Positron computed tomography studies of Chen W, Shi J et al.: Single-impulse panoramic
cerebral glucose metabolism in man: theory and application in photoacoustic computed tomography of small-animal whole-
nuclear medicine. Semin Nucl Med 1981, 11:32-49. body dynamics at high spatiotemporal resolution. Nat Biomed
Eng 2017, 1:71.
6. Ahrens MB, Li JM, Orger MB, Robson DN, Schier AF, Engert F,
Portugues R: Brain-wide neuronal dynamics during motor 21. Tang J, Coleman JE, Dai X, Jiang H: Wearable 3-D
adaptation in zebrafish. Nature 2012, 485:7399 471-U80. photoacoustic tomography for functional brain imaging in
behaving rats. Sci Rep 2016, 6:25470.
7. Boas DA, Brooks DH, Miller EL, DiMarzio CA, Kilmer M,
Gaudette RJ, Zhang Q: Imaging the body with diffuse optical 22. Deán-Ben XL, Sela G, Lauri A, Kneipp M, Ntziachristos V,
tomography. IEEE Signal Processing Magazine 2001, 18:57-75. Westmeyer GG, Shoham S, Razansky D: Functional
8. Ferrari M, Quaresima V: A brief review on the history of human optoacoustic neuro-tomography for scalable whole-brain
functional near-infrared spectroscopy (fNIRS) development monitoring of calcium indicators. Light Sci Appl 2016, 5:e16201.
and fields of application. Neuroimage 2012, 63:921-935.
23. Urban A, Mace E, Brunner C, Heidmann M, Rossier J, Montaldo G:
9. Tanter M, Fink M: Ultrafast imaging in biomedical ultrasound. Chronic assessment of cerebral hemodynamics during rat
IEEE Trans Ultrason Ferroelectr Freq Control 2014, 61:102-119. forepaw electrical stimulation using functional ultrasound
imaging. Neuroimage 2014, 101:138-149.
10. Errico C, Pierre J, Pezet S, Desailly Y, Lenkei Z, Couture O,
Tanter M: Ultrafast ultrasound localization microscopy for 24. Rungta RL, Osmanski B-F, Boido D, Tanter M, Charpak S: Light
deep super-resolution vascular imaging. Nature 2015, 527:7579  controls cerebral blood flow in naive animals. Nat Commun
499-+. 2017, 8:14191.
Using fUS imaging and bi-photon microscopy, this work shows that light
11. Macé E, Montaldo G, Cohen I, Baulac M, Fink M, Tanter M: per se, delivered in trains and at intensities commonly used to trigger
Functional ultrasound imaging of the brain. Nat Methods 2011, functional hyperemia and/or an fMRI signal in rodents, decreases SMC
8:662-664. calcium, either directly or via endothelial cells, leading to dilation of
arterioles. The fact that light also dilates arterioles in the kidney, indicates
12. Berwick J, Johnston D, Jones M, Martindale J, Martin C, that brain specific cells such as astrocytes or microglial cells are not
Kennerley AJ, Redgrave P, Mayhew JEW: Fine detail of players in this effect. The reversibility and reproducibility of ‘photodilation’
neurovascular coupling revealed by spatiotemporal analysis also indicates that it could be used as a simple technical means
of the hemodynamic response to single whisker stimulation in to increase blood flow in a controlled fashion in pathological tissues
rat barrel cortex. J Neurophysiol 2008, 99:787-798. requiring more oxygen.
13. Knecht S, Deppe M, Ebner A, Henningsen H, Huber T, Jokeit H,
25. Tiran E, Ferrier J, Deffieux T, Gennisson JL, Pezet S, Lenkei Z,
Ringelstein E-B: Noninvasive determination of language
Tanter M: Transcranial functional ultrasound imaging in freely
lateralization by functional transcranial doppler sonography: a
moving awake mice and anesthetized young rats without
comparison with the Wada test. Stroke 1998, 29:82-86.
contrast agent. Ultrasound Med Biol 2017, 43:1679-1689.
14. Mace E, Montaldo G, Osmanski BF, Cohen I, Fink M, Tanter M:
Functional ultrasound imaging of the brain: theory and basic 26. Errico C, Osmanski BF, Pezet S, Couture O, Lenkei Z, Tanter M:
principles. IEEE Trans Ultrasonics Ferroelect Frequency Control Transcranial functional ultrasound imaging of the brain using
2013, 60:492-506. microbubble-enhanced ultrasensitive Doppler. Neuroimage
2016, 124:752-761.
15. Van Raaij M, Lindvere L, Dorr A, He J, Sahota B, Foster S,
Stefanovic B: Functional micro-ultrasound imaging of rodent 27. Osmanski BF, Martin C, Montaldo G, Lanièce P, Pain F, Tanter M,
cerebral hemodynamics. Neuroimage 2011, 58:100-108. Gurden H: Functional ultrasound imaging reveals different
odor-evoked patterns of vascular activity in the main
16. Tiran E, Deffieux T, Correia M, Maresca D, Osmanski B-F, olfactory bulb and the anterior piriform cortex. Neuroimage
Sieu L-A, Bergel A, Cohen I, Pernot M, Tanter M: Multiplane wave 2014, 95:176-184.

Current Opinion in Neurobiology 2018, 50:128–135 www.sciencedirect.com

 Functional ultrasound neuroimaging: a review Deffieux et al. 135

28. Gesnik M, Blaize K, Deffieux T, Gennisson JL, Sahel JA, Fink M, 36. Demene C, Bernal M, Delanoe C, Auvin S, Biran V, Alison M,
 Picaud S, Tanter M: 3D functional ultrasound imaging of  Mairesse J, Harribaud E, Pernot M, Tanter M et al.: Functional
the cerebral visual system in rodents. Neuroimage 2017, ultrasound imaging of the brain activity in human neonates.
149:267-274. Sci Transl Med 2017.
This article demonstrates for the first time the ability of fUS imaging to This publication presents the first clinical proof of non-invasive fUS
image in 3D the different brain regions (visual cortex, lateral geniculate imaging of brain activity in humans. Ultrafast ultrasound imaging is
nucleus and superior colliculus) activated during visual stimulation in performed through the fontanel of human neonates at bedside. The
rodents. This results and methodology emphasize the potential of fUS estimation of cerebral blood volume variations in small vessels permits
imaging for advanced neuroscience studies on the visual system in to image the 2D spatiotemporal dynamics of epileptic seizures in com-
rodent models. bination with surface EEG recordings in preterm babies. fUS neuroima-
ging of different sleep phases is also presented in newborns.
29. Rau R, Scheffer W, Belau M, Kruizinga P, De Jong N, Bosch JG,
Maret G: 3D functional ultrasound imaging of the visual system 37. Trevelyan AJ, Baldeweg T, van Drongelen W, Yuste R,
in the pigeon brain. 2017 IEEE International Ultrasonics Whittington M: The source of afterdischarge activity
Symposium (IUS). 2017. in neocortical tonic clonic epilepsy. J Neurosci 2007,
27:13513-13519.
30. Murta T, Leite M, Carmichael DW, Figueiredo P, Lemieux L:
Electrophysiological correlates of the BOLD signal for EEG- 38. Imbault M, Chauvet D, Gennisson J-L, Capelle L, Tanter M:
informed fMRI. Hum Brain Mapp 2015, 36:391-414.  Intraoperative functional ultrasound imaging of human brain
activity. Sci Rep 2017, 7:7304.
31. Urban A, Dussaux C, Martel G, Brunner C, Mace E, Montaldo G:
By imaging small variations of cerebral blood volume in trepanned human
Real-time imaging of brain activity in freely moving rats using
patients undergoing tumor resection, the authors demonstrate for the first
functional ultrasound. Nat Methods 2015, 12:873-878.
time the ability of fUS imaging to image deep brain activity in many
32. Demene C, Bimbard C, Gesnik M, Radtke-Schuller S, Shamma S, different regions of the motor and somatosensorial cortex. It provides a
Boubenec Y, Tanter M: Functional Ultrasound Imaging of the proof of concept of cortical functional mapping based on fUS imaging
thalamo-cortical auditory tract in awake ferrets using ultrafast without electric stimulation. These results demonstrate fUS imaging
Doppler imaging. 2016 IEEE International Ultrasonics Symposium could become a portable and high resolution neuroimaging modality
(IUS). IEEE. 2016:1-4. during brain surgery.

33. Osmanski B-F, Pezet S, Ricobaraza A, Lenkei Z, Tanter M: 39. Bimbar C, Demene C, Girard C, Radtke-Schuller S, Shamma S,
Functional ultrasound imaging of intrinsic connectivity in the Tanter M, Boubenec Y: Multi-scale mapping along the auditory
living rat brain with high spatiotemporal resolution. Nat hierarchy using high-resolution functional UltraSound in the awake
Commun 2014, 5:5023. ferret. bioRxiv 249417, https://doi.org/10.1101/249417
34. Rideau Batista Novais A, Pham H, Van de Looij Y, Bernal M, 40. Kohlhauer M, Lidouren F, Remy-Jouet I, Mongardon N, Adam C,
Mairesse J, Zana-Taieb E, Colella M, Jarreau PH, Pansiot J, Bruneval P, Hocini H, Levy Y, Blengio F, Carli P, Vivien B,
Dumont F et al.: Transcriptomic regulations in oligodendroglial Ricard JD, Micheau P, Walti H, Nadeau M, Robert R, Richard V,
and microglial cells related to brain damage following fetal Mulder P, Maresca D, Demené C, Pernot M, Tanter M, Ghaleh B,
growth restriction. Glia 2016, 64:2306-2320. Berdeaux A, Tissier R: Hypothermic Total Liquid Ventilation Is
Highly Protective Through Cerebral Hemodynamic
35. Knecht S, Dräger B, Deppe M, Bobe L, Lohmann H, Flöel A, Preservation and Sepsis-Like Mitigation After
Ringelstein EB, Henningsen H: Handedness and hemispheric Asphyxial Cardiac Arrest. Crit Care Med 2015,
language dominance in healthy humans. Brain 2000, 123(Pt 43(10:):e420-e430.
12):2512-2518.

www.sciencedirect.com Current Opinion in Neurobiology 2018, 50:128–135